Overcoming Token Limits

Maurice Tony Ewing

20 years: AI and Digital Transformation

In this video, Maurice explains why AI struggles with complex tasks, from data limits to lack of nuance and emotional context. He breaks down token limits, AI’s processing differences from humans, and why human oversight is crucial for accuracy and effectiveness.

Working with AI

Subscribe to watch

Access this and all of the content on our platform by signing up for a 7-day free trial.

Overcoming Token Limits

10 mins 31 secs

Key learning objectives:

Understand AI’s limitations in handling complex, multi-layered tasks due to data dependency, emotional context, and processing constraints
Understand how token limits affect AI’s ability to process, generate, and recall information
Understand why AI struggles with context and nuance, making human oversight essential in collaborative workflows

Skills:

Technology-driven Advancement

Impacts:

Operational Efficiency

Overview:

AI struggles with complex problem-solving due to limitations in data dependency, emotional context, and processing capacity. AI models operate within token limits, restricting how much information they process at once. This can lead to truncated responses, misinterpretations, or an inability to break down complex tasks. While human cognition also has memory constraints, people intuitively prioritise relevant information in ways AI cannot. Understanding AI’s technical limitations helps users manage expectations and optimise collaboration.

Subscribe to watch

Access this and all of the content on our platform by signing up for a 7-day free trial.

Summary

Why can’t AI handle certain complex tasks effectively?

AI struggles with subjective or nuanced decisions, such as selecting aesthetically “ugly but stylish” artwork. It lacks deep cultural awareness, struggles with ethical considerations, and cannot conduct strategic thinking. AI-generated responses are limited by the data it’s trained on, which can lead to biased or fabricated results. Unlike humans, AI cannot break complex tasks into meaningful sub-tasks without guidance, making human input crucial for projects requiring insight, judgment, or emotional intelligence.

What are token limits, and how do they impact AI’s performance?

Token limits define the amount of information an AI model can process at a time. Each word, punctuation mark, or fragment is counted as a token. When a request exceeds the model’s token limit, the AI either truncates the input, leading to incomplete answers, or loses important context. This constraint means AI may miss key details, generate superficial responses, or struggle with multi-step tasks, requiring users to simplify or structure queries effectively.

How does AI process information differently from humans?

AI models recognise patterns in data but do not understand meaning in a human sense. They map relationships based on probability, meaning associations can be skewed if training data is biased or incomplete. AI lacks contextual awareness, emotional intelligence, and intuition. Unlike humans, who prioritise and recall relevant information dynamically, AI’s responses depend entirely on its training data, token capacity, and prompt structure, often requiring human intervention for refinement.

Why is human oversight necessary when working with AI?

AI cannot determine when information is relevant or understand the emotional weight behind interactions. While humans can intuitively adjust communication based on situational awareness, AI applies rigid logic, sometimes missing context or making inappropriate suggestions. Human oversight ensures AI-driven decisions align with strategic goals, ethical considerations, and real-world complexities. Effective AI collaboration requires users to manage expectations, provide structured input, and validate AI-generated outputs before implementation.

Subscribe to watch

Access this and all of the content on our platform by signing up for a 7-day free trial.

Maurice Tony Ewing

Maurice Tony Ewing is a seasoned professional in training, consulting, and executive coaching, having advised and trained thousands of executives and managers across over 60 countries. With a background in banking and board membership, Maurice runs his own consulting firm, Conquer Risk, and teaches data analytics at Columbia University. His most notable accomplishment is leading the data science team that created the world's first mobile-lending platform, which was crucial in Google's creation of "Bebapay" and the NFC payment and lending process. Since then, Maurice has helped companies launch digital initiatives and build strategic, governance, risk, and compliance frameworks. He aims to bring his practical experience and technical knowledge to these learning streams to assist others in learning.

There are no available Videos from "Maurice Tony Ewing"